Aakanksha Agarwal

About Aakanksha Agarwal

Highly analytical ML Scientist with 5+ years of experience architecting, developing, and deploying data-driven solutions that directly impact customer retention, offer optimization, and market analytics. Proven expertise in leading the full cycle of complex projects using ML, LLMs, and Reinforcement Learning to drive significant revenue-generating outcomes and improve key metrics, including a 36% uplift in customer-offer matching while collaborating with cross-functional teams.

Experience

1. Senior Machine Learning Scientist

   Comcast

   Sep 2024 — Present · Philadelphia, PA, US

   Spearheading the Customer Retention Initiative, architecting and optimizing 50+ ML models and post-processing frameworks; leading collaboration with engineering and product teams to define and deliver strategic requirements. Engineered and enhanced offer pricing framework by developing novel division-specific customer profitability features and resolving critical upstream pipeline inconsistencies, achieving a 36% uplift in customer-offer matching. Designed the Next-Best-Action model leveraging 15+ action-propensity models to predict customer preference, including building price-elasticity models and streamlining ML model integration process. Pioneered a Proof-of-Concept for an LLM-based helpdesk solution, processing call transcripts to auto-generate Troubleshooting Guide with potential to reduce average handling time for unresolved calls by 23%. Developed a RAG using KnowledgeGraph (KAG) to build a better understanding of the customer data, integrating Neo4j with Databricks and PySpark for information retrieval using an LLM.

Education

• Birla Institute of Technology and Science, Pilani

  Master's of Science (Honors), Physics

  2014 — 2018

• Columbia University

  Master of Science - MS, Applied Analytics

Skills

• Java
• Microsoft Office
• Research
• Powerpoint
• Microsoft Excel
• Microsoft Word